Overview

The following describes how to upgrade to new versions of Appian on Kubernetes for self-managed customers. This includes upgrades to both the Appian operator and the Appian images.

Scenarios for upgrading

There are different scenarios for when you'd upgrade either the Appian operator or Appian images, or both.

• For major Appian releases, you'll need to upgrade both the Appian operator and Appian images
• For Hotfixes to Appian, you'll typically only need to upgrade the Appian images

Download the new versions and load into Docker

New versions of the Appian images and Appian operator are available for download at:

Support tab on Appian Community > Downloads > Platform > Appian 21.4 > Installers

Depending on what you are upgrading, you can either download the Appian images, the Appian operator (TAR image and TGZ helm chart), or both. See Scenarios for upgrading for more info.

To load the images into Docker, follow the steps under Appian CRD and operator images. Within these steps, depending on what you are upgrading, you may only need to load the operator image, or the Appian images, or both.

Upgrading both the Appian operator and Appian images

To upgrade both the Appian operator and Appian images at the same time, you would:

1. Follow the steps in Upgrading the operator.
2. Follow the steps in Upgrading Appian component images.

Upgrading the operator

The following describes how to upgrade the Appian operator and custom resource definition (CRD) included with the operator install.

With Helm v3, the Appian CRD installed by the operator chart is not updated when running helm upgrade. For more information as to why this is the case, refer to Helm's documentation on custom resource definitions. Instead, the Appian CRD must be extracted from the Appian operator Helm chart and updated manually.

The following steps walk through upgrading the CRD manually, then upgrading the Appian operator.

1. Make sure you've loaded the new version of the Appian operator image into Docker.

2. Untar the Appian operator version you'd like to upgrade to using one of the following commands.

   • If the chart is accessible via a URL, run:

     helm pull <URL> --untar

   • If the chart is stored in a chart repository, run:

     helm pull <CHART_REPO>/appian-operator --version <VERSION> --untar.

   • If you have a copy of the chart locally, but in an archived TGZ file, untar it by running:

     tar -xzf appian-operator-<APPIAN_OPERATOR_VERSION>.tgz.

3. The unpacked Appian operator includes a crds/ directory that contains the CRD. Upgrade the Appian CRD by running:

   kubectl replace -f appian-operator/crds/

4. Update the Appian operator by running helm upgrade:

   1
   2
   3
   4
   helm upgrade appian-operator <CHART: PATH TO TGZ, DIRECTORY, CHART REFERENCE, OR URL> \
   --wait
   --namespace <OPERATOR NAMESPACE> \
   --reuse-values
   

5. Make sure you are now on the new version of the operator by listing the releases:

   helm list -n <OPERATOR NAMESPACE>

6. Make sure the pods of the Appian operator are running:

   kubectl get pods -n <OPERATOR NAMESPACE>

   The controller and webhook pods should both have a STATUS of Running and READY of 1/1.

If you have any trouble upgrading the operator, see Troubleshooting Appian on Kubernetes.

Upgrading the Appian images

The following steps will upgrade the component images of the Appian site including webapp, data-server, and other components. This is typically done when there are new features released.

1. Make sure you've loaded the new version of Appian images into Docker.

2. Stop your Appian site.

   If there are issues shutting down Service Manager when stopping Appian, make sure to shut it down according to Troubleshooting service manager shutdown issues. If Service Manager is not shut down gracefully with these instructions, a subsequent upgrade of Appian could create data corruption.

   If Service Manager can't be brought down gracefully, you'll first need to restart the Appian site and bring it back up on the same version it was already on. Then try to shut down the site again, and make sure everything comes down safely before attempting an upgrade.

3. Within the Appian CR, for each component, update the .spec.<COMPONENT>.image.tag field to be the new version of the image.

4. Restart your Appian site.

If you have any trouble upgrading Appian, see Troubleshooting Appian on Kubernetes.